Nabhas
Trust Model
How we ensure transparency and responsible use of satellite intelligence.
Built on transparency
Open data, explicit timestamps, confidence intervals—no hidden pipelines or black boxes. Trust in satellite intelligence requires knowing what you're looking at, when it was collected, how confident we are, and what can go wrong. Nabhas publishes all of this. We use only open-access satellite data and open-source models. We show confidence intervals, not point estimates. We timestamp everything. We flag limitations explicitly.
Core principles
How we operate
Open sources only
Nabhas uses publicly available satellite data (Sentinel-1, VIIRS, MODIS, ERA5) and open-source models. No classified data, no proprietary satellite feeds.
Baselines, not forecasts
Risk levels are directional signals showing how current observations compare to historical activity. They are not predictions or recommendations.
Timestamps matter
All observations are tied to satellite revisit cadence (typically every few days). Data freshness is explicit. Scores between passes reflect the most recent completed observation.
Example
What satellites actually see
Sentinel-1 B · Kharg Island, Persian Gulf · Three days apart · Same sensor, different signal
Questions
Frequently asked
About Nabhas
What is Nabhas?
A planetary intelligence platform that turns satellite observations into operational state estimates. We combine satellite radar, thermal data, and weather context to infer maritime supply chain risk, wildfire threat, and infrastructure state.
What satellites do you use?
Primarily Sentinel-1 (SAR), VIIRS/MODIS (thermal), and ERA5 (weather). All are open-access data. No classified or proprietary satellite feeds.
How is the data processed?
Each satellite pass goes through a pipeline that normalizes the raw signal, applies machine learning models trained on historical baselines, and produces state estimates with confidence intervals.
Risk levels & confidence
What do risk levels mean?
Low, Moderate, High, and Severe are directional signals—not forecasts. Each reflects how current observations compare to historical activity at that location. A rising score means the zone looks anomalous relative to its own baseline.
Why not just show the raw satellite image?
Raw imagery requires domain expertise to interpret. State estimates with confidence intervals let operators make faster, better decisions. We show evidence on demand, but intelligence comes first.
How confident are these signals?
Every observation includes a confidence score (0-100%). This reflects sensor revisit cadence, cloud cover, model uncertainty, and historical variance. High confidence is rare—be skeptical of 100%.
Data recency
How current is the data?
Observations are tied to satellite revisit cadence. Each zone is updated when a new pass is processed—typically every 2-5 days depending on satellite and location. Scores between passes reflect the most recent completed observation.
Why not real-time updates?
Satellite revisit latency is a physical constraint, not a product choice. Real-time claims over satellite data are misleading. We timestamp everything and show exactly when each observation was collected.
Can I get alerts when things change?
Not yet. Today, Nabhas is a query and discovery tool. Alert systems are planned but require more testing to ensure signal-to-noise ratios are operationally useful.
Limitations & disclaimers
Is this financial or operational advice?
No. Nabhas is an intelligence signal, not a recommendation system. Nothing here constitutes financial, trading, or maritime operational advice.
What can go wrong?
Satellites can be obscured by clouds or equipment failures. Models can misinterpret signal. Baselines can be non-representative during regime shifts. We flag these explicitly when possible.
Who can use Nabhas?
The public dashboard is available to anyone. Sensitive use cases (government, critical infrastructure) require additional vetting. Contact us for partnership conversations.
Data use & privacy
Public access, private API
The dashboard is open to everyone. The API is gated and versioned for trusted partners.
No tracking, no selling
We don't track user behavior or sell data. Usage patterns help us understand demand, but we never publish individual queries.
Export only for non-commercial use
Dashboard data can be downloaded for research, analysis, and internal use. Commercial redistribution requires a partnership agreement.